While artificial intelligence technologies are rapidly transforming workplaces across the private sector, government agencies appear to be adopting these tools at a more cautious pace, potentially missing opportunities to improve efficiency and reduce costs. A recent survey indicates that the public sector, including local and state governments such as Boston and Massachusetts, is lagging in AI utilization compared to private industry.
The nonprofit research group InnovateUS surveyed approximately 100,000 government employees nationwide and found that less than 25 percent reported using AI tools on a weekly basis. This rate is about half that of private-sector workers, where a Gallup poll noted that 45 percent engage with AI regularly. Economists suggest that workers using AI in the private sector save roughly two hours per week, translating into an estimated 1 percent gain in overall U.S. labor productivity.
The slower integration in government settings can partly be attributed to the sector’s focus on sensitive matters such as healthcare, along with challenges related to AI reliability and ethical concerns. Examples include instances where AI-generated legal briefs contained fabricated citations or demonstrated biases in hiring algorithms. Nonetheless, certain businesses have found ways to safely incorporate AI for routine tasks, such as multilingual customer service and automated responses.
In Massachusetts, Governor Maura Healey’s administration made AI tools available to about 40,000 state executive branch employees through a three-year agreement with OpenAI, beginning in February. However, only around 6,500 workers have actively used the tools so far. Chris Smith, director of legislative affairs at the state’s Executive Office of Technology Services and Security, said state employees have employed AI mainly for translating website content and maintaining legacy software. Smith emphasized that employee-driven innovation was a focus of the rollout, enabling workers to identify practical applications of AI in their daily responsibilities.
Boston’s adoption of AI has proceeded more gradually. The city introduced usage guidelines for AI writing and summarization tools roughly three years ago, requiring staff to complete training before accessing Gemini, Google’s AI assistant. In addition to Gemini, Boston deploys AI to optimize traffic light timing and facilitate public queries of municipal data. Shin-pei Tsay, the city’s chief information officer, expressed optimism about AI’s potential benefits while stressing the need for careful implementation.
Experts highlight that training and clear policies are critical for successful AI adoption in government roles. Beth Simone Noveck, founder of InnovateUS and former chief innovation officer for New Jersey, noted that many government employees have not yet used AI due to uncertainty about permitted applications and lack of familiarity with the technology. Among non-users surveyed, 28 percent said they did not know how AI could apply to their jobs, 23 percent lacked skills to use the tools, and 19 percent were unsure about permission to engage with AI.
Some states are actively using AI to improve public services. For example, Nevada employs AI to reduce errors in administering the Supplemental Nutrition Assistance Program (SNAP) and to assist entrepreneurs in completing necessary business forms. Gary Blank, a senior fellow at the Pioneer Institute in Boston who studies government AI, advocates for expanded adoption in Massachusetts and beyond. He noted that the discourse often swings between extremes—either fearing AI's existential risks or ignoring its practical benefits—and urged governments to focus on realistic applications that enhance efficiency without undue alarm.
As AI continues to evolve, public sector agencies face the challenge of balancing innovation with caution to realize gains in productivity while maintaining public trust and security.